在线优化参数的多变量非线性预测函数控制  

Multivariable nonlinear predictive function control based on parameter optimization online

作　　者：侯小秋 Hou Xiaoqiu(School of Electrical&Control Engineering,Heilongjiang University of Science&Technology,Harbin 150022,China)

机构地区：[1]黑龙江科技大学电气与控制工程学院,哈尔滨150022

出　　处：《黑龙江科技大学学报》2023年第3期439-444,共6页Journal of Heilongjiang University of Science And Technology

摘　　要：针对实用随机多变量NARMAX模型的控制问题,利用多变量非线性递推最小二乘法进行参数估计,使用在动态工作点处的动态切平面逼近该模型,转化为时变多变量CARMAX模型,由其构成递推预测模型进行预测,基于多变量CARMAX模型的预测函数控制,运用非线性递推最小二乘法优化控制输入输出的加权系数,给出在线优化加权系数的多变量非线性预测函数控制算法。结果表明,系统的控制响应具有优良的性能,无超调和稳偏,调节时间约5个采样周期。This paper aims to improv the control function of multivariable practical random NARMAX model.The study involves estimating the parameters by using the multivariable nonlinearity recursive least squares method;approaching to the model at working point with a dynamic cutting horizontal a method and translating into time-varying multivariable CMRMAX model as the recuisive prediction model;based on the predictive function control of non-linear multivariable CARMAX model,employing the nonlinearity recursive least squares method to optimize controlling in-put and out-put weighting factor,and obtaining an multivariable nonlinear predictive function control algorithm with online optimization.The simulation results show that the control response of the system exhibits excellent performance without overshoot or stable deviation;and the settling time is about 5 sampling periods.

关 键 词：预测函数控制 非线性控制 随机多变量NARMAX模型 非线性递推最小二乘法 在线优化参数 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]